How to tune waiting for logflush

Using Singlestore Cloud managed service.

Porting an application doing a high amount of concurrent bulk inserts (among other things and yes I know it’s not the best way to load data). My problem is that 40% of total stmnt time is spent in “Waiting for logflush” and for the leaf nodes it’s >85%.

Fairly new to Singlestore so how can I tune the log flush and IO in general for the Cloud version?